Ivanov & Filanovsky - OpAmp Speed and Accuracy Improvement, Kluwer, 2004
The improvement of speed and accuracy in typewritting
Transcript of The improvement of speed and accuracy in typewritting
LB
7Z<^. 7
^CHANGE
THE IMPROVEMENT OF SPEED AND
ACCURACY IN TYPEWRITING
A dissertation submitted to the Board of
University Studies of the Johns Hopkins
University in conformity with the require-
ments for the degree of Doctor of
Philosophy
1922
BY
ROY EDWARD HOKE
BALTIMORETHE JOHNS HOPKINS PRESS
1922
of'
THE IMPROVEMENT OF SPEED ANDACCURACY IN TYPEWRITING
THE IMPROVEMENT OF SPEED AND
ACCURACY IN TYPEWRITING
A dissertation submitted to the Board of
University Studies of the Johns Hopkins
University in conformity with the require-
ments for the degree of Doctor of
Philosophy
1922
BY
ROY EDWARD HOKE
BALTIMORETHE JOHNS HOPKINS PRESS
1922
&otoer* printing CompanpLEBANON, PA.
t^iAjjiX
TABLE OF CONTENTSJ
Chapters Page
Introduction 9
I. The Frequency of the Occurrence of Letters and
Marks in the English Language 11
II. Errors in Typewriting and Their Causes 20
III. Relative Abilities of the Eight Fingers and the
Two Hands for Typewriting 24
IV. The Finger and Hand Loads of the Present Type-
writer Keyboard 29
V. The Consideration to be Taken Into Account in
the Arrangement of an Ideal Keyboard Based
on Scientific Principles, and a Suggestion for
Such Keyboard 31
Bibliography 42
Vita 43
t) v'_> \f v ) /<£ O
INTRODUCTION
Some time ago the writer began work on the problem of
constructing standard tests in typewriting, such as have been
and are being formulated for the various subjects of the ele-
mentary and secondary school curricula. The first prelimi-
nary investigations, however, revealed the fact that there
were other typewriting problems which might logically claim
priority. First, since typewriting is a matter entirely of the
use of the letters of the alphabet and the various marks, it
seemed necessary to investigate the frequency of the occur-
rence of these letters and marks in the English language.
Secondly, since accuracy is one of the primary requisites of
good typewriting, it appeared to be important to investigate
the number and distribution of errors made in actual type-
writing, and to learn, if possible, their causes. Third, the
modern "touch" method of typewriting presumably makes
equal demands upon all the fingers and upon the two hands.
In fact, the typewriter keyboard is so arranged as to assign
fifteen letters to the left hand, and only eleven to the right.
The problem arises : what are the relative abilities of the
eight fingers and the two hands. Fourth, the determination
of the loads or burdens of work which the present typewriter
keyboard places upon the fingers and hands. Fifth, an at-
tempt to enumerate the considerations which must be taken
into account in the arrangement of a keyboard according to
scientific principles; a suggested keyboard, and the criticism
of the same from the considerations set forth.
The field of this investigation is one that has not been
cultivated. With the exception of the work by Book on
"The Psychology of Skill," little or no experimentation has
been done in typewriting. Even in Book's study, the object
was not so much the improvement of speed, accuracy or
methods of teaching typewriting, as it was the more general
aim of ascertaining, by the use of the typewriter as a mere
9
10 INTRODUCTION
bit of apparatus, the psychology of skill. Any other mechan-
ism, an adding machine or piano for instance, would have
been equally capable of use in a study of skill, though per-
haps not equally convenient for experimental purposes.
I. THE FREQUENCY OF THE OCCURRENCE OFLETTERS AND MARKS IN THE ENGLISH
LANGUAGE
The first objective of the problem under consideration is
to determine the frequency with which the various letters
of the alphabet and the more common marks are used in
the English language. For this purpose valuable material
was readily at hand as to the letters in the form of the Ayres
Spelling Scale. This scale itself informs us that the 1,000
words in the list are "the product of combining different
studies with the object of identifying the 1,000 commonest
words in English writing." The scale was formed by com-
bining four different vocabulary studies as follows: (1)
A study by Ayres himself involving a total of 23,629 words
found in personal and business letters. (2) The study of
Eldridge comprising 43,989 words found on two pages each
of four different newspapers. (3) The study of Rev.
Knowles of London, of the vocabulary of the English Bible
and various authors, to a total of 100,000 words. (4) The
study of Cook and O'Shea of the vocabulary used by thirteen
persons in family and social correspondence, including a
total of 5,200 different words. Most of these words occurred
a number of times. It will be seen, therefore, that the
1,000 words in the Ayres Scale are those that were found
most frequently in approximately 368,000 words of personal
letters, business letters, newspapers, the Bible, and various
English authors.
Accepting the 1,000 words of this list as the most commonin the English language, it would appear that by counting
the frequency with which the various letters of the alphabet
appear in these words, we should secure a very close approxi -
mation to their frequency in the English language in general,
and in the work done on typewriters in particular.
11
12 THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH
Table I shows in the first column the gross number of
occurrences of each letter of the alphabet in the 1,000 com-
monest words. A total of 5,433 letters were found. The
average length of the words, therefore, is 5.433 letters. Bymultiplying the number of occurrences of each letter by
TABLE I
Frequency of Use of Letters in the Words of the AyresSpelling Scale
THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH 13
Although the evidence thus secured for the frequencies
of letters seemed almost conclusive, it was nevertheless
thought desirable to support it by corroborative evidence
from original studies. With this purpose in view, three
studies were made on the frequencies of letters and punctua-
TABLE II
Frequency of Use of Letters and Punctuation Marks in theSixteenth Chapter of Mark's Gospel
14 THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH
raised to 1,000 for purposes of comparison. This was done
by multiplying the first column through by 3.69.
Table III shows the frequencies of use of letters and
punctuation marks in actual business letters. The first col-
TABLE III
Frequency of Use of Letters and Punctuation Marks in Busi-ness Letters
Times Used Raised to iooo
THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH 15
A comparison of the data of these three original studies
with Table I, which gives the frequency of letters only in
the Ayres Spelling Scale, with the letter "E" raised to 1,000
for purposes of comparison, will reveal much similarity both
TABLE IV
Frequency of Use of Letters and Punctuation Marks in aNewspaper Editorial
16 THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH
of correlation is .946 and the probable error is .012. The
product-moment formula was used in securing this correla-
tion.
It is of course possible that in any one study of the fre-
quency of letters the chance presence or absence of a few
words containing the less common letters, such as K, J, Q,
X or Z, may appreciably influence the result. This difficulty
TABLE VStudies on the Frequency of Use of Letters and Punctuation
Marks
THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH 17
while in the fifth column of the table we have the average
of the first four. This column is based on the counting of
12,130 letters of the alphabet in connection with the four
studies. The findings set forth in this fifth column
are accepted in this study as a basis for evaluating
the various letters and marks and their importance or fre-
quency in typewriting.
TABLE VI
Frequency of Use of Letters and Punctuation MarksPrinter's Fonts as Compared With Averages
Given in Table V
IN
Printer*
18 THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH
Rg.l Graphic Representation of the
Frequency of the Use of Letters.
THE FREQUENCY OF LETTERS AND MARKS IN ENGLISH 19
Figure 1 shows graphically the frequency of the various
letters. The fact is that the first six letters (E, T, A, O, S, I)
are used more frequently than the remaining twenty letters.
The most startling observation is the fact that "E" is used
more frequently than twelve other letters combined (Y, B,
P, W, F, G, V, K, J, X, Q, Z). In view of these facts the
conclusion is very evident, viz., that the frequencies of the
letters must be taken into consideration and form the basis
for any scientific arrangement of the typewriter keyboard.
In further corroboration of the findings of this study as
to the frequencies of letters and marks, the custom of the
printing trade may be adduced. Table VI represents in the
first column the number of type of each letter included in
a large printer's font, the same for comparative purposes
being reduced to the basis of 1,000 for "E" by dividing
through by 12, and the results shown in the second column.
Our accepted findings are given in the third column. Even
a casual comparison of these data reveals so close an agree-
ment between them that a correlation by statistical methods
would be superfluous. It needs, however, to be borne in
mind that the printer's font supplies a somewhat larger
number of the less common letters and marks than would
be usually found necessary. Furthermore, the fact that in
the printer's font four letters (O, S, I, N) are given the same
rank, with 8,000 type each, indicates that we have here only
a rough approximation of usage, and that the evidence from
the make-up of the font is valuable for our purpose for
corroboration only and not for the definite evaluation of
letters.
II. ERRORS IN TYPEWRITING AND THEIRCAUSES
For the purposes of this part of our investigation about
500 full size pages of practice typewriting work were secured,
from approximately 100 different individuals, and the errors
found therein counted and tabulated up to the point where
"E" was charged with 1,000 errors. This point was reached
when all but three pages of the material had been checked.
The errors for the several letters are given in Table VII,
TABLE VII
Errors Made in Actual Typewriting as Compared With Fre-quency of Occurrence of Letters
ERRORS IN TYPEWRITING AND THEIR CAUSES 21
tionship between these two series of data is given in the
third column of Table VII, from which it will be seen that
twelve letters are as good or better than "E" from the stand-
point of accuracy, while thirteen rank below "E." The
median is 108.3.
The relationship between accuracy and frequency of use
may be definitely calculated by the use of the product-moment
formula. The coefficient of correlation is found to be .924
TABLE VIII
Comparison of Frequency of Use of Letters With Accuracy
22 ERRORS IN TYPEWRITING AND THEIR CAUSES
thirteen most used letters, in the lower half the thirteen
least frequent letters. In the first column the frequency of
use of letters is indicated by a plus sign for those letters
above the median in use, and a minus sign for those
below the median. In the second column a plus sign indi-
cates those letters above the median in number of errors,
and a minus sign those below the median in number of
errors. There are only two of the thirteen most commonletters which are below the median in number of errors,
and then by a very narrow margin. Also only two of the
thirteen least used letters are above the median in number
of errors, nor then by very large margins. In the third
column the plus sign indicates those letters which are above
the median in percentage of accuracy, while the minus sign
indicates those letters which are below the median in accu-
racy. With five exceptions out of the 26 letters, frequency
of use and accuracy go together, or infrequency of use with
inaccuracy. One of the five exceptions is by a negligible
difference.
We had expected to find that some keys, or rows of keys,
on the typewriter would be more accurate than others. The
findings contradict this expectation, and indicate that the
position of keys on the present keyboard exercises very little
if any measurable effect upon accuracy. It has not been
found possible to say with any degree of assurance that any
one position on the keyboard is better than any other from
the standpoint of accuracy.
A further study has been made to ascertain whether the
errors made in typing any letter are more likely to occur
in combination with certain letters than with others. Toanswer this question errors were tabulated until 100 had
been noted for "E," and in each case the letter after which
the error occurred was noted. The results of this study are
set forth in Table IX. This table should be read as follows
:
ERRORS IN TYPEWRITING AND THEIR CAUSES 23
TABLE IX
Errors and the Letters After Which They Occurred
E M20, L14, R12, N12, S6, T6, V5, H5, Z5, K3, F3, B3, 02, G2.T 010, A9, 18, S8, N4, C4, E3, F3, B3, U3, 03.R E16, 010, A10, M4, H3, P3.
A H10, T10, 04, W4, E3, D3, R2, C2, F2, M2, SI, VI.S E10, A12, 14, R4, U5, T2, 02, N2, B2.N N9, A8, 08, R3, U2, 12.
L8, R6, Y5, D4, T4, G2, H2, 02, E2, W2.1 H6, L5, P4, W2, M2, V2, R2, K2, A2, N2.D E12, 08, N4, L2, V2, Rl.L U6, P6, E4, F2, 02, A2, 12, G2, Rl, Bl.G Nil, E4, A4, T2, Rl, Ol.H W6, P5, C4, T4, G2, Al.M 06, 14, U2, E2, L2, R2, SI.
U 06, R4, B3, Y2, D2, Gl, HI.Y A6, N4, B3, T2, F2, R2, N2, H2, CI.C E8, A4, 12, N2, S2, Ul.V A5, 14, 04, N3, E2, Rl.W A8, S7.F A3, E2, 02, Nl.B R2, 12, M2, Ol, Ul.K 02, R2, 12, Nl, Al.P H3, 12, Ul, El.X A3, 02, 12, El.
Q E2.
J A2.Z D2, Al.
Tabulating from the above the errors made after each of the let-
ters we get the following: After E 81 errors, A 81, O 75, N 46,
R 45, I 34, H 32, L 31, M 30, T 30, S 25, U 22, P 18, B 15, W 14,
F 12, C 11, D 11, V 10, G 7, Y 7, K 5, Z 5, J 0, Q 0, X 0. Inspection
shows that this correlates very highly with the frequency of use of
letters, which would indicate merely that most errors happen after the
letters most frequently used because most trials occur, and therefore
most chance for error after the frequently used letters.
of the 100 errors made in writing "E," 20 occurred after M,14 after L, 12 after R, 12 after N, 6 after S, 6 after T,
5 after V, 5 after H, 5 after Z, 3 after K, 3 after F, 3 after
B, 2 after O, and 2 after G. It does not appear from a
study of these data that there is any connection between
accuracy and the combination of letters with which any one
letter may happen to be written. This serves only to con-
firm the conclusion that accuracy or inaccuracy is the
result of frequent or infrequent practice, due to frequent
or infrequent use of the letter.
III. RELATIVE ABILITIES OF THE EIGHTFINGERS AND THE TWO HANDS FOR
TYPEWRITING
For the purpose of determining the relative abilities of the
eight fingers and the two hands three studies were made.
The first of these studies was made upon 50 High School
girls; the second on 46 High School boys. In those two
studies the following method was used. The subjects were
instructed to hold the thumb of the right hand against the
center of the typewriter frame, in front of the space bar, and
to tap the letter "J" w^n tne first finger of the right hand
as rapidly as possible. Thirty seconds were allowed. Then
with the thumb of the left hand in the same position, the
subject tapped with the first finger of the left hand the letter
"F" for thirty seconds. Then in the same way, with the
right hand second finger "K," left hand second finger "D,"
right hand third finger "L," left hand third finger "S,"
right hand fourth finger" ;" left hand fourth finger "A,"
each for thirty seconds. The number of taps leaving their
impressions on typewriter paper in the regular way were
counted for each of the eight fingers of each of the 96
subjects.
The purpose of holding the thumb against the metal frame
of the typewriter was to eliminate, so far as possible, wrist
and arm movements, for it was desired to test only finger
abilities. It was found that by holding the thumb in this
position the subject was compelled to make the taps by relying
upon the fingers alone, as was desired for the purposes of
the experiment. The results of these two studies are set
forth in Tables X and XI.
24
RELATIVE ABILITIES OF FINGERS AND HANDS 25
TABLE XTapping Test on Typewriter—High School Girls
Showing Number of Taps Made by Each Finger in 30 Seconds
26 RELATIVE ABILITIES OF FINGERS AND HANDS
TABLE XI
Tapping Test on Typewriter—High School BoysShowing Number of Taps Made by Each Finger in 30 Seconds
Number ofSubject
Left HandFingers used
I3 | 2 |
Letters struck
Is | d |
Right HandFingers usedI
2 | 3I 4
Letters struck
i Ik | 1
1
2345
67
8910
11
12
13
14
15
16
17
18
19
2021
22232425
262728293031
32333435363738394041
4243444546
AveragesMedians
157124149
105
150
135
122129137
129
138
139103
126
94150
93181
134112
129112124
139100121
148
90123
129
114119
105
109
134113
121
137129
151
142
140
118159
124
130
169
142
154108
152
150138
158130121
143
140126172
79146
115
186
138|
143j
129|
128|
117
131
122
134139
101!
110|
134160
124117
152
163
145
163
193
151
138
141
125
155
96153
136186158
126
143
131
124141
|
139|
151!
142j
120|
128 I
131
184163
123
154
145
126
175
161
135
140155
121
148
94157
139
176
168
144
145
122
132
138
142
147165
108
133
131
RELATIVE ABILITIES OF FINGERS AND HANDS 27
An inspection of the medians in these two tables will
show that the eight fingers have about the same rank and
relative ability for the two sexes. The only noticeable sex
difference is in speed of tapping, with respect to which the
boys excelled the girls 24.6 per cent.
The third study was made of 42 college girls and 12
teachers of Frederick County, Maryland. In this study a
somewhat different method was pursued. The subjects were
required to tap for one minute with a pencil with the right
hand, then one minute with the left hand. The marks were
subsequently counted to get the scores. (A later study
showed that when the subjects attempted to count their taps
as they went, their speed was impeded. A few reported that
this served to increase their scores.) The subjects were
further required to tap for thirty seconds upon their desks
with each finger. Particular instructions were given as to
the position to be assumed : with the wrist resting upon the
desk (to prevent wrist movement in tapping) and with three
fingers also resting with their tips upon the desk. It wasfound that this position effectually prevented any wrist or
arm movement being used and secured records of the tapping
abilities of the fingers alone.
An inspection of the medians for the eight fingers in Table
XII shows that they very closely bear out the results of
the first and second studies as shown in Tables X and XI.
The three studies above involve a total of 190,410 taps,
made by 150 different individuals. It is believed that by
combining these three studies by the method of averages,
we may secure conclusions as to the relative abilities of the
fingers and hands which may be relied upon.
28 RELATIVE ABILITIES OF FINGERS AND HANDS
TABLE XII
Tapping Test—42 College Girls, 12 Teachers
Number ofSubject
Taps per min.
Right LeftHand Hand
Taps in 30 Seconds
RlILI || R2
IL2 R3 L3 || R4
IL4
1
23
45
67
8910
11
12
13
14
15
16
17
18
19
202122232425262728293031
3233343536373839
4041
424344454647
48495051
5253
54
355360446357443386382368324326
I
390
1 387
I
409
I
397
I
417
!334
I
398
I
359
I
363
I
342
I342
i400
I
352
I
326343294383367355322253374407309415354363340434392317281
290402332370
3061
291
367318340350314304175
254328312299273307299318314286331329378330372271
289309282265285291314310226334303326346344288205280292360340300
360| 280400| 380404| 342408 392374| 314314| 374408| 298472| 324
185
IV. THE FINGER AND HAND LOADS OF THEPRESENT TYPEWRITER KEYBOARD
Since the foregoing studies make it possible to state in a
mathematical form the relative abilities of the two hands and
the eight fingers, and the relative values of each letter and
mark, it would seem worth while first to raise the question,
what is being expected of the fingers and hands by the pres-
ent typewriter keyboard. We are assuming, of course, that
30 LOAD OF THE TYPEWRITER KEYBOARD
In accordance with the touch method each of the eight
fingers has its own proper keys, three for each finger except
the first finger of each hand which has six. Table XIII
gives for each finger the letters it strikes and their values,
and the sum of these, which is the finger load. These letter
values are taken from Table V. Denoting the fingers of the
right hand as Rl, R2, R3 and R4, and those of the left hand
as LI, L2, L3 and L4, we find their loads to be, in the
same order: Rl, 1,490; R2, 640; R3, 996; R4, 296; LI,
1,535; L2, 1,492; L3, 658; L4, 803. Adding these finger
loads for each hand we get the hand loads : for the right hand
3,422, for the left 4,488. In view of the fact that in Tables
X, XI and XII we found finger abilities to be not very
dissimilar, it is rather startling to find that Rl, LI, and L2are each, on the present typewriter keyboard, given more
than five times as much work to do as is given to R4. In
view, also, of the generally known fact that the index of
right-handedness may be roughly expressed by the ratio of
ten to nine, that is, that the ability of the right hand is
approximately one-ninth greater than that of the left, it is
surprising to find that the present keyboard gives a far
heavier load to the weaker member.
The question may be raised: Why are 100 strikes re-
corded for both the right and left hand Shift Key? This
was gotten by actual count, 200 being recorded. However,
whether the right hand strikes exactly half of them or not
will vary somewhat with the nature of the matter to be
written. When a letter on the left side of the present key-
board is capitalized the right shift key must be struck and
vice versa. Inasmuch as most of the letters are on the left
side of the keyboard and many of them the more frequent
in use, it is very probable that the right shift key is the more
frequently used. However, the total number will remain
the same and an equal division of the total 200, while not
entirely exact, is as near as can be gotten.
V. THE CONSIDERATIONS TO BE TAKEN INTOACCOUNT IN THE ARRANGEMENT OF ANIDEAL KEYBOARD BASED ON SCIEN-TIFIC PRINCIPLES, AND A SUG-
GESTION FOR SUCH KEY-BOARD
The several considerations which should control a scientific
rearrangement of the typewriter keyboard are the following
:
1. The measured abilities of the eight fingers and of the
right as compared with the left hand, should be taken into
consideration, and loads assigned in proportion to strength.
2. The measured frequency of use of the letters of the
alphabet and punctuation marks must be made the basis for
calculating the loads to be assigned to the several fingers.
3. In addition to the above there are certain pedagogical
considerations
:
(a) The most used letters should be assigned to the
"home" or guide keys, since on a keyboard so arranged the
work could be done with the fewest possible changes of
position of hands. Also the frequent use of these letters will
aid the pupil in the early fixing of these home-key positions.
(b) The next most used letters should be assigned to
those keys or positions which appear to be the favorable ones
from the standpoint of accuracy.
(c) Since beginners find it easier to use the first fingers
than the other fingers, and since the approved method of
teaching typewriting, as well as other subjects, is that of
proceeding from the easy to the more difficult, it is desirable
so to arrange the keyboard, by assigning several of the
vowels to the first fingers, that simple practice words and
sentences may be written by the beginner, even in the earliest
stages of the learning process.
For purposes of the first consideration, Table XIV as-
sembles the data from Tables X, XI and XII, i. e., the
median tapping abilities of the eight fingers of high school
31
32 AN IDEAL KEYBOARD
TABLE XIVComparison of Finger and Hand Abilities With the Loads of
the Fingers and Hands
Left Hand Right Hand
I 3 I4 Total
Mds. for H. S. Girls..! 106 I 119Mds. for H. S. Boys..| 127
|134
Mds. College Girls and|Teachers | 107
|103
Averages 1 113|
119
117145
12s129
119145
145136
119156
160145
130166
138145
126I
117164 151
104131
117128=1046
Multiplying these avs.by 7.56 to get idealtypewriting loads:....
The loads on the pres-ent keyboard are....
Percentage of over- orunder-load of the fin-
gers on present type-writer keyboard
855I
900I
803I
658I
—6.1I—26.9
975 1028 1097 1096 968
1492I 1535 II 1490
I640
I 996 I296
1+53- I +49.3 1 1 +35.8I—41.6 +.5 1—69.4
Left Hand Right HandTapping ability (190,410 taps by 150
individuals) 89,593 taps 100,817 tapsRatio: 88.87 to 100
Present load 4,488 taps 3,422 tapsRatio: 131.25 to 100
Per cent of over- or under-load 47.7
(over-load of left
hand as comparedwith right.)
girls, high school boys, and college girls and teachers, com-
bining the three by the method of averages and showing a
total of finger abilities of 1,046. Table XIV also includes
the present keyboard load from Table XIII. Dividing the
total keyboard load (7,910—the sum of 3,422 and 4,488 of
Table XIII) by the sum of the finger abilities (1,046, Table
XIV) we find that each point of finger ability must be multi-
plied by 7.56 so as to reapportion the whole load upon the
fingers in proportion to their abilities. Comparing the ideal
finger loads with the present finger loads, we find four over-
loads of 49.3, 53., 35.8, and .5; also four under-loads, 5.1,
26.9, 41.6, and 69.4.
Of the 190,410 taps made in this study, 100,817 were madeby the right hand, and 89,593 by the left, in equal time.
Therefore the ratio of the ability of the right hand to that
of the left is as 100 to 88.87, or approximately as ten is to
nine. This agrees with the findings of Bryan.* The present
*Brvan, American Journal of Psychology, 1892, V, 123-204.
AN IDEAL KEYBOARD 33
right and left hand loads in typewriting are 3,422 taps for
the right to 4,488 for the left, or a ratio of 100 to 131.25.
These data show an overload of the left hand of 47.7 per
cent as compared with the right hand, on the present type-
writer keyboard. The over-loads and under-loads discov-
ered are more graphically shown in Figure 2.
The criticism might perhaps be advanced that in combining
the measures of the tapping abilities of the several fingers
as obtained on the typewriter, with those obtained in tapping
on desks, two different, and for all we know, heterogeneous
sets of data are being added together. In reply to this sug-
gestion the following considerations are advanced and need
to be taken into account.
First, it should be borne in mind that the same purpose
was uppermost in these two sorts of experimentation, namely,
the attempt to get a mathematical statement of the com-
parative abilities of the several fingers. The two experi-
ments are alike in that both attempt to secure an accurate
count of the number of taps each finger can make in the
same period of thirty seconds, or of one minute.
Second, it is true that tapping on a desk is not, so far as
we know, the same thing as tapping the keys of a typewriter
in the ordinary use of that machine. But it is also true that
tapping the keys of a typewriter under the controlled situa-
tion, namely, with the thumb touching the frame of the
machine, beneath the space bar, may, for all we know, be
no more like the process of actual typewriting than is tapping
on a desk as described above. In fact, it may be that the
former more nearly approximates the actual process involved
in typewriting than does the latter. Both are controlled
situations, with the purpose of preventing an undue influ-
ence upon the outcome as a result of leaving the arm and
wrist free to do the tapping, instead of compelling the fingers
to do it.
Third, of the three sets of data which are combined in
Table XIV, the third (the results of tapping on desks) does
not differ more from either the first (the results for high
34 AN IDEAL KEYBOARD
N *> *
*3
N K> ^-
« «c
J5b
§3
to o
1si
*
AN IDEAL KEYBOARD 35
school girls tapping on typewriters) or the second (the
results for high school boys tapping on typewriters) than
the first differs from the second. Therefore, an argument
against including the third, because it differs from the first
and second, would be just as good an argument, so far as it
goes, for excluding either the first or second because it differs
from the other.
Fourth, the average scores for the fingers secured by com-
bining the three sets of data in Table XIV differ from the
scores which would have been secured by combining only
the two first sets of data of that table by only a little over
four per cent. So far as the relative abilities of the two
TABLE XVComparison of Finger-loads of Standard and Proposed Keyboards
With Ideal Loads
36 AN IDEAL KEYBOARD
Fifth, the fact is generally known, especially among
pianists and typists, that the third finger is the least capable.
This fact is not so clearly revealed in the first two sets of
data of Table XIV as in the third. This may perhaps be
explained by saying that the third finger, in the case of the
experiment on the typewriter, being further removed from
the fixed point, the thumb, than was the first or second finger,
was perhaps aided somewhat in its tapping by the movement
TABLE XVI
The Accuracy of Horizontal Lines of Keys on Typewriter
Upper
AN IDEAL KEYBOARD 37
data in the averages gives a truer estimate of the abilities
of the fingers than would be the case were these data omitted.
To return to the considerations which must be taken into
account in arranging a keyboard along scientific lines, it was
noted as one of the pedagogical considerations that the eight
most frequently used letters should be assigned to the home
keys, so far as this may be found to be possible without
overloading of fingers. Also, that several of the vowels
should be assigned to the first fingers, for the reason stated.
TABLE XVIIThe Accuracy of the Eight Fingers in Typewriting
(Percentages)
right handFingers 1 2 3 4
o 46y 141
38 AN IDEAL KEYBOARD
of keys. In Table VIII we show that, with reference to
the individual letters, accuracy and frequency of use are
almost perfectly correlated. Table XVI, however, seems to
show that we have the greatest accuracy in the upper hori-
zontal line, with decreasing accuracy in the middle and
especially in the lower lines, to an extent that is not justified
by use or disuse. Since this is the case, it will be well to
assign the more frequently used letters, after the home keys
have been supplied, to the upper row of keys.
The second part of Table XVI, on the other hand, shows
differences in accuracy of fingers, that is, of vertical rows
of keys. The accuracy data are taken from Table XVII.
But while the accuracy of fingers seems to differ, it appears
from the table that this is due to the loads assigned to the
fingers and hands, rather than to qualities of the fingers
or hands themselves. For when the fingers are given their
rank for accuracy and for load, we find a close correlation.
The overloaded fingers tend to be inaccurate. In the case
of the two hands the evidence is even more clear. The right
hand with the smaller load has the better accuracy score.
The average frequency of use of letters written by the left
hand on the present keyboard is 291, by the right 280. So
far as the law of use goes, the left hand should be four
per cent more accurate than the right. Instead of this we
find the right hand 84% more accurate than the left. This
must be due to the over-load of the left hand.
This evidence tends to show that nearly half of the errors
made by the left hand in typewriting are preventable, since
they are due to the bad apportioning of the typewriting load
on the present keyboard.
A further important reason why the typewriter keyboard
should be scientifically rearranged is the self-evident fact
that maximum speed and ease of operation can never be
attained so long as some fingers are over-worked, while
others do not have a chance to contribute their full share
to the total result.
AN IDEAL KEYBOARD 39
A rearrangement of the keyboard will, it is clear, make
for improved speed and accuracy in typewriting.
The present standard arrangement of the typewriter key-
board antedates the "touch" method of operating, which has
come to be the only accepted method in the teaching of type-
writing. Therefore the considerations, if any, which sug-
gested the present arrangement of the keyboard, do not apply,
but new considerations must be brought to bear which take
account of the fact that all of the fingers are to be used
and should contribute their share, no more and no less.
In accordance with the above considerations we offer the
preferred rearrangement of the letters and marks of the
typewriter keyboard shown in Figure 3.
In addition to the improvements in the arrangement of
the typewriter keyboard based on the investigations and
considerations hereinbefore set forth, Figure 3 shows certain
other changes from the conventional arrangement which
remain to be explained. These changes are (1) the inter-
posing between the right and left banks of keys of the shift
lock, shift, and back spacer keys, and (2) the placing of
the margin and tabular keys in the lower horizontal row of
keys, one at either end.
In general, it may be said, first of all, that both these
changes are made in an effort to make the keyboard arrange-
ment more compact, thereby bringing all keys within reach
of the fingers without unnecessary hand movements. These
keys are, for the most part, usually so placed on the type-
writer as to require the operator to remove the fingers from
the letter keys. Furthermore, the shift key, shift lock, and
back spacer are the heaviest keys on the typewriter. The back
spacer operates against the tension of the machine and is prob-
ably the heaviest key of all. The shift key and shift lock
usually lift either the carriage or the basket of keys. In
either case, the operation of these keys requires far more
force than the operation of the letter keys. It is felt, there-
fore, that these three keys should be placed in a central
position, as in Figure 3, so that they may be operated by the
40 AN IDEAL KEYBOARD
AN IDEAL KEYBOARD 41
thumb or first finger, rather than by the fourth finger, whose
strength is scarcely adequate. The margin release and tabu-
lar key are very light in touch, and not often used, so that
in assigning them to the fourth fingers, as in Fig. 3, we are
not unduly burdening these weak fingers.
The advantages of this arrangement are: 1. The keys in
question are brought into a convenient position for operation
by the thumb or first finger of either hand. 2. The entire
keyboard can be operated without removing the hands from
the position they take when the fingers are placed upon the
"home keys," as the keys of the middle horizontal row of
letters are called in touch typewriting. 3. The interposition
of the shift, shift-lock and back spacer keys divides the
keyboard into right and left halves, thereby tending to avoid
errors and undesirable hand movement resulting from the
encroachment of one hand on the territory of the other. 4.
The interposed keys accentuate the importance and make
necessary the confining of each hand to its appropriate bank
of keys, with a minimum of movement for complete perform-
ance of the entire duty assigned it. These four advantages,
together with all the other advantages claimed for the key-
board as shown in Figure 3, may be summed up, in a general
way, by saying that this keyboard is definitely planned for the
most accurate and economical operation by the touch method,
the method which is now being taught in practically all
schools, whereas the present conventional, so-called standard
and universal keyboards have been arranged with reference
to no discoverable or known criteria whatsoever.
BIBLIOGRAPHYAyres, Leonard P. (1915.) Measuring Scale for Ability in Spell-
ing. Russell Sage Foundation, New York.(1912). The Spelling Vocabularies of Personal and Busi-
ness Letters. Russell Sage Foundation, New York.Book, W. F. (1908). The Psychology of Skill: With Special
Reference to Its Acquisition in Typewriting. University of
Montana Publications in Psychology; Bulletin No. 53. Psycho-logical Series No. 1.
Bryan, W. L. (1892). On the Development of Voluntary MotorAbility. American Journal of Psychology, Vol. 5, pp. 123 ff.
Bryan and Harter. Studies in Telegraphic Language. Psycho-logical Review, Vol. 6.
Burk, F. (1898). From Fundamental to Accessory in the Develop-ment of the Nervous System and of Movements. PedagogicalSeminary, Vol. 6, pp. 5-64.
Cook and O'Shea (1914). The Child and His Spelling. Indian-
apolis, Ind.
Eldridge (1911). Six Thousand Common English Words. NiagaraFalls, N. Y.
Fritz and Eldridge. Expert Typewriting. American Book Com-pany, New York.
Gordon, Kate (1917). Educational Psychology. Henry Holt andCompany, New York.
Gould, G. M. (1908). Right-handedness and Left-handedness. J. B.
Lippincott Company, Philadelphia.
Judd, C. H. (1915). The Psychology of the High School Subjects.
Ginn and Company, New York.Knowles, Rev. J. (1904). The London Point System of Reading
for the Blind. London, England.Owen, Margaret B. (1917). The Secret of Typewriting Speed.
Forbes and Company, Chicago.Parker, S. C. (1915). Methods of Teaching in High Schools. Ginn
and Company, New York.Ross, J. W. (1921). Typewriting Manual. Rowe Publishing Com-
pany, Baltimore, Md.SoRelle, Rupert P. and Cutler, Ida McL. Rational Typewriting.
The Gregg Publishing Company, New York.SoRelle, Rupert P. Methods of Teaching Typewriting. The Gregg
Publishing Company, New York.Starch, D. Periods of Work in Learning. Journal of Educational
Psychology, Vol. 3, pp. 209-213.
Thorndike, E. L. Practice in the Case of Typewriting. Pedagogi-cal Seminary, Vol. 20, pp. 516-529.
(1914). Educational Psychology: Briefer Course. TeachersCollege, Columbia University.
Wilson, D. (1891). The Right Hand: Left-handedness. The Mac-millan Company, New York.
Washburn, M. F. (1916). Movement and Mental Imagery. Hough-ton Mifflin Company, New York.
Woodworth, R. S. (1903). Le Mouvement. Paris.
Accuracy of Voluntary Movement. Psychological ReviewMonograph 3.
Woolley, H. B. The Development of Right-handedness in a Nor-mal Infant. Psychological Review, 17. 1910, pp. 37-41.
VITA
Roy Edward Hoke, son of Edward John and MaryMargaret Hoke, born at York, Pennsylvania, July 3, 1896.
Grammar school education was received in Lancaster, Pa.,
Lewistown, Pa., and Newark, N. J. College preparatory
course was taken in the Boys' High School, Lancaster, Pa.
Attended Franklin and Marshall College, Lancaster, Pa.,
1912 to 1916, received the A.B. degree. The A.M. degree
was received in the following year, 1917. Was elected a
member of the Phi Beta Kappa Society in the Theta Chapter
upon graduation in 1916. Taught during the school year
1916-1917 in the Columbia Grammar School, New YorkCity. During the following year attended the Theological
Seminary of the Reformed Church in the United States, at
Lancaster, Pa. Enlisted in the U. S. Army May 30, 1918,
serving as Assistant Chaplain with the rank of Sergeant, at
Fort Oglethorpe, Ga., receiving honorable discharge Janu-
ary 13, 1919. Attended Johns Hopkins University during
the years 1919-1920, 1920-1921. Ordained into the Chris-
tian ministry of the Reformed Church August 14, 1919,
and installed Pastor of the Reformed Charge, Jefferson,
Md. At the University courses were pursued in Education,
Psychology and Philosophy, under the instruction of Prof.
Edward F. Buchner, Prof. Knight Dunlap, Prcf. A. O.
Lovejoy, Dr. David E. Weglein, Dr. Henry Slonimsky, and
Dr. Florence E. Bamberger, to each of whom the writer
is indebted, and hereby expresses his appreciation.